Iain Sandoe [Sun, 26 May 2019 19:35:14 +0000 (19:35 +0000)]
darwin, x86, testsuite - Amend three PIE tests,
Darwin requires PIC code in order to support PIE, amend the test scan-asms
to match this.
2019-05-26 Iain Sandoe <iain@sandoe.co.uk>
* gcc.target/i386/pr39013-1.c: Adjust scan-asms for PIE to
account for PIC code on Darwin.
* gcc.target/i386/pr39013-2.c: Likewise.
* gcc.target/i386/pr64317.c: Likewise.
Iain Sandoe [Sat, 25 May 2019 19:42:05 +0000 (19:42 +0000)]
darwin, x86, testsuite - Check for an expected error on Darwin.
The test requests an alignment which exceeds the maximum object
file aligment for Darwin, rather than skipping it - test that we
see the expected error.
2019-05-25 Iain Sandoe <iain@sandoe.co.uk>
* gcc.target/i386/pr89261.c: Test that the alignment required
by the test correctly produces the expected error on Darwin.
re PR target/90530 (Invalid SUBREG insn generated by reload)
PR target/90530
* config/pa/pa.c (pa_can_change_mode_class): Accept mode changes from
DImode to SImode in floating-point registers on 64-bit target.
* config/pa/pa.md (umulsidi3): Change nonimmediate_operand to
register_operand in xmpyu patterns.
Andrew Stubbs [Fri, 24 May 2019 15:25:41 +0000 (15:25 +0000)]
GCN: Wait for exit value to write before exiting.
2019-05-24 Andrew Stubbs <ams@codesourcery.com>
gcc/
* config/gcn/gcn-run.c (main): Set a non-zero return value if the
kernel does not exit cleanly.
* config/gcn/gcn.md (gcn_return): Insert s_waitcnt before s_dcache_wb.
Richard Biener [Fri, 24 May 2019 11:31:11 +0000 (11:31 +0000)]
re PR middle-end/90607 (gcc.dg/pr53265.c FAILs)
2019-05-24 Richard Biener <rguenther@suse.de>
PR testsuite/90607
* tree-loop-distribution.c (struct partition): Add location
member.
(partition_alloc): Initialize all fields.
(generate_memset_builtin): Use the location recorded in the
partition for the generated call.
(generate_memcpy_builtin): Likewise.
(classify_partition): Record the location of a single store
as location for the partition.
* gcc.dg/pr53265.c: Amend for new expected diagnostic.
Commit r271514 missed changing the type of two functions in
aarch64-protos.h. The function definitions had been updated to use
uint64_t while the function declarations had been missed.
They were missed since I only tested the patch on aarch64 where
`unsigned long` is the same as `uint64_t`.
This patch updates these declarations in aarch64-protos.h.
Tested by building an aarch64 cross-compiler on arm-none-linux-gnu (so
that `unsigned long` and `uint64_t` are different and would give error
messages), and bootstrapping on aarch64-none-linux-gnu.
Also manually tested command line options to see that
-march=armv8-a+typo prints out the expected flags while using the new
feature flags does not complain about missing flags.
gcc/ChangeLog:
2019-05-24 Matthew Malcomson <matthew.malcomson@arm.com>
PR target/90588
* common/config/aarch64/aarch64-common.c
(aarch64_rewrite_selected_cpu): Change local temporary variable
type from unsigned long to uint64_t.
* config/aarch64/aarch64-protos.h (aarch64_parse_extension,
aarch64_get_extension_string_for_isa_flags): Change declaration to
match new definition by replacing unsigned long with uint64_t.
Jakub Jelinek [Fri, 24 May 2019 08:59:37 +0000 (10:59 +0200)]
re PR libgomp/90585 (libgomp hsa plugin ftbfs in the x32 multilib variant)
PR libgomp/90585
* plugin/plugin-hsa.c: Include gstdint.h. Include inttypes.h only if
HAVE_INTTYPES_H is defined.
(print_uint64_t): New typedef.
(PRIu64): Define if HAVE_INTTYPES_H is not defined.
(print_kernel_dispatch, run_kernel): Use PRIu64 macro instead of
"lu", cast uint64_t HSA_DEBUG and fprintf arguments to print_uint64_t.
(release_kernel_dispatch): Likewise. Cast shadow->debug to uintptr_t
before casting to void *.
* plugin/plugin-nvptx.c: Include gstdint.h instead of stdint.h.
* oacc-mem.c: Don't include config.h nor stdint.h.
* target.c: Don't include config.h.
* oacc-cuda.c: Likewise.
* oacc-host.c: Don't include stdint.h.
Jakub Jelinek [Fri, 24 May 2019 08:58:50 +0000 (10:58 +0200)]
re PR target/90568 (stack protector should use cmp or sub, not xor, to allow macro-fusion on x86)
PR target/90568
* config/i386/x86-tune-sched.c (ix86_macro_funsion_pair_p): Call
gen_attr_type just once instead of 4-7 times. Formatting fixes.
Handle stack_protect_test_<mode> codegen similarly to corresponding
sub instruction.
Richard Biener [Fri, 24 May 2019 08:48:14 +0000 (08:48 +0000)]
re PR tree-optimization/88440 (size optimization of memcpy-like code)
2019-05-23 Richard Biener <rguenther@suse.de>
PR tree-optimization/88440
* opts.c (default_options_table): Enable -ftree-loop-distribute-patterns
at -O[2s]+.
* tree-loop-distribution.c (generate_memset_builtin): Fold the
generated call.
(generate_memcpy_builtin): Likewise.
(distribute_loop): Pass in whether to only distribute patterns.
(prepare_perfect_loop_nest): Also allow size optimization.
(pass_loop_distribution::execute): When optimizing a loop
nest for size allow pattern replacement.
Iain Sandoe [Fri, 24 May 2019 07:24:26 +0000 (07:24 +0000)]
Darwin, x86, testsuite - Make match strings more specific.
Some of the i386.exp tests fail on Darwin (and at least one passes
incorrectly on Linux) because their scan-asm match strings are too
general. In some cases the strings also match instances in the .file
and size directives or in comment output. This patch makes the match
strings more specific.
2019-05-24 Iain Sandoe <iain@sandoe.co.uk>
* gcc.target/i386/pconfig-1.c: Scan for the string in the generated
code, not in comments or miscellaneous directives. Adjust expected
count.
* gcc.target/i386/pr18041-1.c: Likewise.
* gcc.target/i386/pr18041-2.c: Likewise.
* gcc.target/i386/wbinvd-1.c: Likewise.
* gcc.target/i386/wbnoinvd-1.c: Likewise.
* gcc.target/i386/pr66819-3.c: Specifically, check that there is no
call to "bar".
* gcc.target/i386/pr66819-4.c: Likewise.
* gcc.target/i386/pr82662.c
* gcc.target/i386/ptwrite2.c: Make the checks look for the specific
destination register, don't try the m32 test on m64 targets.
Iain Sandoe [Fri, 24 May 2019 07:05:25 +0000 (07:05 +0000)]
Darwin, x86, testsuite - Disable tests that cannot pass.
A few of the i386.exp target tests cannot pass on Darwin
because either the port doesn't support a feature, or the
ABI demands sufficiently different codegen that matching
the output with scan-asms would require a different test.
Jonathan Wakely [Thu, 23 May 2019 21:41:08 +0000 (22:41 +0100)]
Fix testsuite bugs
One of the static assertions in 20_util/function_objects/invoke/1.cc was
wrong, but didn't fail because by default it was compiled with
-std=gnu++14 which didn't use that static assertion. Split out the C++17
parts to a new file that always runs with -std=gnu++17, so those checks
are always done.
The 23_containers/unordered_set/allocator/ext_ptr.cc test is supposed to
be a run-time test but was unintentionally compile-only.
* testsuite/20_util/function_objects/invoke/1.cc: Move C++17-specific
tests to ...
* testsuite/20_util/function_objects/invoke/3.cc: New test.
* testsuite/23_containers/unordered_set/allocator/ext_ptr.cc: Change
"compile" test to "run".
* doc/xml/manual/evolution.xml: Document LWG DR 2921 change.
* doc/xml/manual/intro.xml: Likewise.
* include/std/future (__create_task_state): Add default arguments
to make providing an allocator optional.
(packaged_task::packaged_task(F&&)): Call __create_task_state directly
instead of delegating to another constructor.
(packaged_task::packaged_task(allocator_arg_t, const A&, ...)): Do not
define allocator-extended constructors for C++17 and later.
* testsuite/30_threads/packaged_task/cons/alloc.cc: Only run test for
C++11 and C++14.
* testsuite/30_threads/packaged_task/cons/alloc2.cc: Likewise.
* testsuite/30_threads/packaged_task/cons/alloc_min.cc: Likewise.
* testsuite/30_threads/packaged_task/uses_allocator.cc: Likewise.
Iain Sandoe [Thu, 23 May 2019 20:51:04 +0000 (20:51 +0000)]
darwin, x86 - disable '-mfentry' for the port.
This can't work without special support because placing
the call directly after the function label breaks the
ABI requirement that call sites are 16byte aligned.
Bill Schmidt [Thu, 23 May 2019 20:33:02 +0000 (20:33 +0000)]
rs6000.c (rs6000_global_entry_point_needed_p): Rename to rs6000_global_entry_point_prologue_needed_p.
[gcc]
2019-05-22 Bill Schmidt <wschmidt@linux.ibm.com>
* config/rs6000/rs6000.c (rs6000_global_entry_point_needed_p):
Rename to rs6000_global_entry_point_prologue_needed_p. Return
false for PC-relative functions.
(rs6000_output_function_prologue): Change called function name to
rs6000_global_entry_point_prologue_needed_p. Emit ".localentry
name,1" for PC-relative functions.
(rs6000_elf_declare_function_name): Change called function name to
rs6000_global_entry_point_prologue_needed_p.
From what I understand of the libstdc++/83237 thread at
<https://gcc.gnu.org/ml/gcc-patches/2017-12/msg00573.html>, the
high numbers are not arbitrary, so it seems wrong to try
lowering them, or we'd just waste cycles testing nothing, or
worse, ending up with a bogus error indication. Better to just
plain disable this part of the test for simulator targets; I
assume the results should be the same on any IEEE-float target,
i.e. no target-specific things going on here that'd raise a need
to cover it everywhere.
With this part of the test disabled, I saw the test finishing in
(time) "124.74s user" where it was before "1120.26s user"
running the cris-elf-run simulator on a "i7-4770K CPU @ 3.50GHz"
host. Most certainly that indidates that the remainder of the
test is still too much for *some* host+simulator combos, but I'm
happy with the runtime lowered to 1/5 of the timeout (10
minutes) on this particular combination, and I'd think this
fixes timeouts for many other simulator combos too.
This construct (disabling or lowering limits for simulators) is
used elsewhere in the libstdc++ test-suite and in particular the
SIMULATOR_TEST macro is used in the testsuite machinery (though
AFAICT not in testDiscreteDist).
* testsuite/26_numerics/random/poisson_distribution/operators/values.cc:
Don't run the libstdc++/83237 part on simulator targets.
Jan Hubicka [Thu, 23 May 2019 16:07:07 +0000 (18:07 +0200)]
re PR tree-optimization/90576 (SPEC CPU2006 450.soplex miscompiled with -Os -flto after r271413)
PR tree-optimization/90576
* tree-ssa-alias.c (compare_sizes): Remove dead calls to
poly_int_tree_p.
(aliasing_component_refs_p): Fix three way size compare conditional;
give up earlier in case we can not decide on equivalence.
Co-Authored-By: Martin Liska <mliska@suse.cz>
From-SVN: r271572
Jonathan Wakely [Thu, 23 May 2019 14:13:18 +0000 (15:13 +0100)]
Make any_cast compare typeinfo as well as function pointers
It's possible for the function pointer comparison to fail even though
the type is correct, because the function could be defined multiple
times with different addresses when shared libraries are in use.
Retain the function pointer check for the common case where the check
succeeds, but compare typeinfo (if RTTI is enabled) if the first check
fails.
* include/experimental/any (__any_caster): Use RTTI if comparing
addresses fails, to support non-unique addresses in shared libraries.
* include/std/any (__any_caster): Likewise.
Jonathan Wakely [Thu, 23 May 2019 13:39:06 +0000 (14:39 +0100)]
PR libstdc++/90220 fix experimental::any_cast for non-object types
This corresponds to the fixes done for std::any_cast, but has to be done
without if-constexpr. The dummy specialization of _Manager_internal<_Op>
is used to avoid instantiating the real _Manager_internal<T>::_S_manage
function just to compare its address.
PR libstdc++/90220
* include/experimental/any (__any_caster): Constrain to only be
callable for object types. Use remove_cv_t instead of decay_t.
If the type decays or isn't copy constructible, compare the manager
function to a dummy specialization.
(__any_caster): Add overload constrained for non-object types.
(any::_Manager_internal<_Op>): Add dummy specialization.
* testsuite/experimental/any/misc/any_cast.cc: Test function types
and array types.
Martin Liska [Thu, 23 May 2019 12:07:09 +0000 (14:07 +0200)]
Do not use tree_to_wide_ref that point to a temporary (PR c++/90587).
2019-05-23 Martin Liska <mliska@suse.cz>
PR c++/90587
* tree-ssa-uninit.c (value_sat_pred_p): The result of &
operation points to a temporary (pointed via tree_to_wide_ref)
that is out of scope after the &.
Richard Biener [Thu, 23 May 2019 11:35:16 +0000 (11:35 +0000)]
re PR tree-optimization/88440 (size optimization of memcpy-like code)
2019-05-23 Richard Biener <rguenther@suse.de>
PR tree-optimization/88440
* opts.c (default_options_table): Enable -ftree-loop-distribute-patterns
at -O[2s]+.
* tree-loop-distribution.c (generate_memset_builtin): Fold the
generated call.
(generate_memcpy_builtin): Likewise.
(distribute_loop): Pass in whether to only distribute patterns.
(prepare_perfect_loop_nest): Also allow size optimization.
(pass_loop_distribution::execute): When optimizing a loop
nest for size allow pattern replacement.
There was a regression for gfortran.dg/fmt_en.f90 for cris-elf that on...
There was a regression for gfortran.dg/fmt_en.f90 for cris-elf
that on inspection was due to it having acquired a truncation
call through the runtime. I updated that and the new tests that
had "Fortran runtime error: required ftruncate or chsize support
not present" messages in gfortran.log, ran past cris-elf and
committed as obvious. See also
<https://gcc.gnu.org/ml/gcc-patches/2008-05/msg00975.html> from
which I copy-pasted most of this message. (Yep, 11 years ago.)
gcc/testsuite:
* gfortran.dg/dec_io_1.f90, gfortran.dg/dtio_1.f90,
gfortran.dg/dtio_12.f90, gfortran.dg/fmt_en.f90,
gfortran.dg/namelist_89.f90: Gate test on effective_target
fd_truncate.
David Malcolm [Thu, 23 May 2019 00:42:03 +0000 (00:42 +0000)]
Bulletproof -fdiagnostics-format=json against bad locations (PR c++/90462)
PR c++/90462 reports an ICE with -fdiagnostics-format=json when
attempting to serialize a malformed location to JSON.
The compound location_t in question has meaningful "caret" and "start"
locations, but has UNKNOWN_LOCATION for its "finish" location,
leading to a NULL pointer dereference when attempting to build a JSON
string for the filename.
This patch bulletproofs the JSON output so that attempts to write
a JSON object for a location with a NULL file will lead to an object
with no "file" key, and attempts to write a compound location with
UNKNOWN_LOCATION for its start or finish will lead to the corresponding
JSON child object being omitted.
This patch also adds a json::object::get member function, for self-testing
the above.
gcc/ChangeLog:
PR c++/90462
* diagnostic-format-json.cc: Include "selftest.h".
(json_from_expanded_location): Only add "file" key for non-NULL
file strings.
(json_from_location_range): Don't add "start" and "finish"
children if they are UNKNOWN_LOCATION.
(selftest::test_unknown_location): New selftest.
(selftest::test_bad_endpoints): New selftest.
(selftest::diagnostic_format_json_cc_tests): New function.
* json.cc (json::object::get): New function.
(selftest::test_object_get): New selftest.
(selftest::json_cc_tests): Call it.
* json.h (json::object::get): New decl.
* selftest-run-tests.c (selftest::run_tests): Call
selftest::diagnostic_format_json_cc_tests.
* selftest.h (selftest::diagnostic_format_json_cc_tests): New
decl.
gcc/testsuite/ChangeLog:
PR c++/90462
* g++.dg/pr90462.C: New test.
Jonathan Wakely [Wed, 22 May 2019 22:14:34 +0000 (23:14 +0100)]
PR libstdc++/90557 fix path assignment that alters source
PR libstdc++/90557
* src/c++17/fs_path.cc (path::_List::operator=(const _List&)): Fix
reversed arguments to uninitialized_copy_n.
* testsuite/27_io/filesystem/path/assign/copy.cc: Check that source
is unchanged by copy assignment.
* testsuite/util/testsuite_fs.h (compare_paths): Use std::equal to
compare path components.
Jason Merrill [Wed, 22 May 2019 21:39:08 +0000 (17:39 -0400)]
PR c++/20408 - unnecessary code for empty struct.
Here initializing the argument from a TARGET_EXPR isn't an empty class
copy even though the type is !TREE_ADDRESSABLE, so we should check
simple_empty_class_p.
* call.c (build_call_a): Use simple_empty_class_p.
Jonathan Wakely [Wed, 22 May 2019 20:29:39 +0000 (21:29 +0100)]
PR libstdc++/77691 fix resource_adaptor failures due to max_align_t bugs
Remove the hardcoded whitelist of allocators expected to return memory
aligned to alignof(max_align_t), because that doesn't work when the
platform's malloc() and GCC's max_align_t do not agree what the largest
fundamental alignment is. It's also sub-optimal for user-defined
allocators that return memory suitable for any fundamental alignment.
Instead use a hardcoded list of alignments that are definitely supported
by the platform malloc, and use a copy of the allocator rebound to a POD
type with the requested alignment. Only allocate an oversized
buffer to use with std::align for alignments larger than any of the
hardcoded values.
For 32-bit Solaris x86 do not include alignof(max_align_t) in the
hardcoded values.
PR libstdc++/77691
* include/experimental/memory_resource: Add system header pragma.
(__resource_adaptor_common::__guaranteed_alignment): Remove.
(__resource_adaptor_common::_Types)
(__resource_adaptor_common::__new_list)
(__resource_adaptor_common::_New_list)
(__resource_adaptor_common::_Alignments)
(__resource_adaptor_common::_Fund_align_types): New utilities for
creating a list of types with fundamental alignments.
(__resource_adaptor_imp::do_allocate): Call new _M_allocate function.
(__resource_adaptor_imp::do_deallocate): Call new _M_deallocate
function.
(__resource_adaptor_imp::_M_allocate): New function that first tries
to use an allocator rebound to a type with a fundamental alignment.
(__resource_adaptor_imp::_M_deallocate): Likewise for deallocation.
* testsuite/experimental/memory_resource/new_delete_resource.cc:
Adjust expected allocation sizes.
* testsuite/experimental/memory_resource/resource_adaptor.cc: Remove
xfail.
Jason Merrill [Wed, 22 May 2019 19:48:05 +0000 (15:48 -0400)]
PR c++/86485 - simple_empty_class_p
Yet another tweak that would have fixed this bug: we should treat INIT_EXPR
and MODIFY_EXPR differently for determining whether this is a simple empty
class copy, since a TARGET_EXPR on the RHS is direct initialization if
INIT_EXPR but copy if MODIFY_EXPR.
* cp-gimplify.c (simple_empty_class_p): Also true for MODIFY_EXPR.
H.J. Lu [Wed, 22 May 2019 18:53:37 +0000 (18:53 +0000)]
x86: Don't allocate stack frame nor align stack if not needed
get_frame_size () returns used stack slots during compilation, which
may be optimized out later. This patch does the followings:
1. Add stack_frame_required to machine_function to indicate that the
function needs a stack frame.
2. Change ix86_find_max_used_stack_alignment to set stack_frame_required.
3. Always call ix86_find_max_used_stack_alignment to check if stack
frame is needed.
Tested on i686 and x86-64 with
--with-arch=native --with-cpu=native
Tested on AVX512 machine configured with
--with-arch=native --with-cpu=native
gcc/
PR target/88483
* config/i386/i386-options.c (ix86_init_machine_status): Set
stack_frame_required to true.
* config/i386/i386.c (ix86_get_frame_size): New function.
(ix86_frame_pointer_required): Replace get_frame_size with
ix86_get_frame_size.
(ix86_compute_frame_layout): Likewise.
(ix86_find_max_used_stack_alignment): Changed to void. Set
stack_frame_required.
(ix86_finalize_stack_frame_flags): Always call
ix86_find_max_used_stack_alignment. Replace get_frame_size with
ix86_get_frame_size.
* config/i386/i386.h (machine_function): Add stack_frame_required.
gcc/testsuite/
PR target/88483
* gcc.target/i386/stackalign/pr88483-1.c: New test.
* gcc.target/i386/stackalign/pr88483-2.c: Likewise.
This patch adds support in the compiler for the architecture feature
flags that binutils will use to enable/disable the new "Future
Architecture Technologies" feature Scalable Vector Extension V2 (SVE2)
announced at Linaro Connect this week.
The "sve2" extension that enables the core sve2 instructions.
This also enables the sve extension, since sve is a requirement of sve2.
Extra optional sve2 features are the bitperm, sm4, aes, and sha3 extensions.
These are all given extra feature flags, "bitperm", "sve2-sm4",
"sve2-aes", and "sve2-sha3" respectively.
The sm4, aes, and sha3 extensions are explicitly marked as sve2
extensions to distinguish them from the corresponding NEON extensions.
When introducing macros to denote these new features we have gone past
what a 32 bit value can represent which means we need to change the type
of those variables working with these feature flags to ensure they use
64 bit quantities.
Tested with bootstrap on aarch64-none-linux-gnu and manually seeing that
-march=armv8-a+typo prints out the expected flags while using the new
feature flags does not complain about a missing flag (until reaching the
assembler).
gcc/ChangeLog:
2019-05-22 Matthew Malcomson <matthew.malcomson@arm.com>
* common/config/aarch64/aarch64-common.c
(struct aarch64_option_extension, struct processor_name_to_arch,
struct arch_to_arch_name, aarch64_parse_extension, opt_ext_cmp,
aarch64_contains_opt,
aarch64_get_extension_string_for_isa_flags): Change type of
variables storing flags to uint64_t.
* config/aarch64/aarch64-option-extensions.def (sve2, sve2-sm4,
sve2-aes, sve2-sha3, bitperm): New optional SVE2 extension flags.
* config/aarch64/aarch64.c (struct processor,
aarch64_parse_arch, aarch64_parse_cpu, aarch64_validate_mcpu,
aarch64_validate_march, aarch64_override_options,
aarch64_option_print, aarch64_handle_attr_isa_flags,
aarch64_declare_function_name, aarch64_start_file): Make flag
variables uint64_t.
* config/aarch64/aarch64.h (AARCH64_FL_SVE2, AARCH64_FL_SVE2_AES,
AARCH64_FL_SVE2_SM4, AARCH64_FL_SVE2_SHA3,
AARCH64_FL_SVE2_BITPERM): New macro feature flags.
* config/aarch64/aarch64.opt (aarch64_isa_flags): Make uint64_t.
* config/aarch64/driver-aarch64.c
(struct aarch64_arch_extension, struct aarch64_core_data,
struct aarch64_arch_driver_info, host_detect_local_cpu): Make
flag variables uint64_t.
* doc/invoke.texi: Add documentation for new arguments.
Andrew Stubbs [Wed, 22 May 2019 13:12:03 +0000 (13:12 +0000)]
Fix fortran size_type_node parameter passing.
2019-05-22 Andrew Stubbs <ams@codesourcery.com>
gcc/fortran/
* trans-stmt.c (gfc_trans_critical): Use size_type_node for
gfor_fndecl_caf_lock and gfor_fndecl_caf_unlock calls.
(gfc_trans_allocate): Use size_type_node for gfor_fndecl_caf_sync_all
call.
Janne Blomqvist [Wed, 22 May 2019 11:56:01 +0000 (14:56 +0300)]
fortran/89100: Default widths with -fdec-format-defaults
gcc/fortran ChangeLog:
2019-05-22 Jeff Law <law@redhat.com>
Mark Eggleston <mark.eggleston@codethink.com>
PR fortran/89100
* gfortran.texi: Add Default widths for F, G and I format
descriptors to Extensions section.
* invoke.texi: Add -fdec-format-defaults
* io.c (check_format): Use default widths for i, f and g when
flag_dec_format_defaults is enabled.
* lang.opt: Add new option.
* options.c (set_dec_flags): Add SET_BITFLAG for
flag_dec_format_defaults.
gcc/testsuite ChangeLog:
2019-05-22 Mark Eggleston <mark.eggleston@codethink.com>
PR fortran/89100
* gfortran.dg/fmt_f_default_field_width_1.f90: New test.
* gfortran.dg/fmt_f_default_field_width_2.f90: New test.
* gfortran.dg/fmt_f_default_field_width_3.f90: New test.
* gfortran.dg/fmt_g_default_field_width_1.f90: New test.
* gfortran.dg/fmt_g_default_field_width_2.f90: New test.
* gfortran.dg/fmt_g_default_field_width_3.f90: New test.
* gfortran.dg/fmt_i_default_field_width_1.f90: New test.
* gfortran.dg/fmt_i_default_field_width_2.f90: New test.
* gfortran.dg/fmt_i_default_field_width_3.f90: New test.
libgfortran ChangeLog:
2019-05-22 Jeff Law <law@redhat.com>
PR fortran/89100
* io/format.c (parse_format_list): set default width when the
IOPARM_DT_DEC_EXT flag is set for i, f and g.
* io/io.h: add default_width_for_integer, default_width_for_float
and default_precision_for_float.
* io/write.c (write_boz): extra parameter giving length of data
corresponding to the type's kind.
(write_b): pass data length as extra parameter in calls to
write_boz.
(write_o): pass data length as extra parameter in calls to
write_boz.
(write_z): pass data length as extra parameter in calls to
write_boz.
(size_from_kind): also set size is default width is set.
* io/write_float.def (build_float_string): new paramter inserted
before result parameter. If default width use values passed
instead of the values in fnode.
(FORMAT_FLOAT): macro modified to check for default width and
calls to build_float_string to pass in default width.
(get_float_string): set width and precision to defaults when
needed.
Richard Biener [Wed, 22 May 2019 07:44:24 +0000 (07:44 +0000)]
re PR tree-optimization/90450 (Hash function in gather_mem_refs_stmt does not match with mem_ref_hasher::equal)
2019-05-22 Richard Biener <rguenther@suse.de>
PR tree-optimization/90450
* tree-ssa-loop-im.c (struct im_mem_ref): Add ref_decomposed.
(mem_ref_hasher::equal): Check it.
(mem_ref_alloc): Initialize it.
(gather_mem_refs_stmt): Set it.
Alan Modra [Wed, 22 May 2019 04:34:26 +0000 (14:04 +0930)]
[RS6000] Don't pass -many to the assembler
I'd like to remove -many from the options passed by default to the
assembler, on the grounds that a gcc bug in instruction selection (eg.
emitting a power9 insn for -mcpu=power8) is better found at assembly
time than run time.
For now, just do this when --enable-checking or gcc is not a release.
This patch also emits .machine assembler directives for ELF targets
when functions are compiled for different cpus via attributes or
pragmas. That's necessary when the initial -m<cpu> option passed to
the assembler doesn't enable the superset of all opcodes emitted, as
seen by a failure of gcc.target/powerpc/clone2.c without .machine
when building gcc for power8.
rs6000_machine_from_flags deliberately uses ISA_2_4_MASKS rather than
ISA_2_2_MASKS for power5 because "friz" and other similar instructions
enabled by gcc with TARGET_FPRND are enabled in gas by "-mpower5".
(gas -mpower5 supports power5+ too.) rs6000-cpus.def puts
OPTION_MASK_FPRND in ISA_2_4_MASKS, so ISA_2_4_MASKS is the one to use
in deciding to pass "-mpower5" to gas.
O3-pr70130.c also failed on an earlier version of this patch (when
only testing one ISA bit to determine .machine). This is a test for a
power7 vector bug, but on power8 hw check_vect_support_and_set_flags
passes -mpower8-vector which means the test isn't exercising the
original bug exactly. I reckon that is wrong, and similary for other
vector testcases that ask for a specific cpu. I've fixed it here by
explicitly passing -mno-power8-vector and similar vector options.
* config/rs6000/rs6000.h (ASM_OPT_ANY): Define.
(ASM_CPU_SPEC): Conditionally add -many.
* config/rs6000/rs6000.c (rs6000_machine): New static var.
(rs6000_machine_from_flags, emit_asm_machine): New functions..
(rs6000_file_start): ..extracted from here, and modified to
test all ISA bits.
(rs6000_output_function_prologue): Emit .machine as necessary.
* testsuite/gcc.target/powerpc/ppc32-abi-dfp-1.c: Don't use
power mnemonics.
* testsuite/gcc.dg/vect/O3-pr70130.c: Disable default options
added by check_vect_support_and_set_flags.
* testsuite/gcc.dg/vect/pr48765.c: Likewise.
* testsuite/gfortran.dg/vect/pr45714-b.f: Likewise.
In <https://gcc.gnu.org/ml/gcc-patches/2019-05/msg01417.html>...
In <https://gcc.gnu.org/ml/gcc-patches/2019-05/msg01417.html>:
...a change introduced in r94104 where BITS_PER_WORD was chosenperhaps because we expect register-sized writes into this area.
See <https://gcc.gnu.org/ml/gcc-patches/2005-01/msg01564.html>
for the r94104 change.
* function.c (assign_parm_setup_block): Raise alignment of
stacked parameter only for STRICT_ALIGNMENT targets.
projects / thirdparty / gcc.git / log
